Predictive frequency-based sequence estimator for control of grid-tied converters under highly distorted conditions

This paper proposes a novel frequency-based predictive sequence extractor that allows to isolate the harmonic components of both voltages and currents needed for the control of grid-tied converters. The proposed method is based on a modification of the Sliding Goertzel Transformation (SGT) that allows to include a predictive behavior with a prediction horizon equal to the processing window needed for the algorithm. The technique performance is compared with the well-established DSOGI alternative, allowing for a higher bandwidth in the estimation as well as improved immunity to changes in the magnitude, frequency and phase of the tracked signals. Additionally, the impact of the proposed method on the closed-loop performance of the current controlled converter is proposed as a metric, thus enabling other researches to have a clear view about the expected real impact of the different existing methods.

I. Introduction

Distributed power generation (DPG) is expected to play an important role in the short and medium term design of the generation, transport and distribution system. This is due to the penetration of renewable generation units that allows to produce power, providing at the same time ancillary services (harmonic compensation [1], magnitude and frequency restoration [2],…) An engaging characteristic of the DPG systems based in renewable generation is that they help to decrease the emissions since the DPG units are placed near the power is consumed. On the other hand, the use of DPG increases the complexity of the whole system due to the coexistence of several systems with different characteristics (nominal power, output impedance, workload, transient response …)